10,000 datasets for each, 1-star and 5-star data from amazon reviews.

Contains:
- Source texts
- Doc2vec 50 dim embeddings
- 2 dim embeddings computed with umap
- Minimal python example on how to access data